energies

Article
A Non-Uniform Transmission Line Model of the
±1100 kV UHV Tower
Xianshan Guo 1 , Ying Fu 1 , Jingqiu Yu 2, * and Zheng Xu 2
1 State Grid Corporation of China, Beijing 100000, China; xianshang-guo@sgcc.com.cn (X.G.);
ying-fu@sgcc.com.cn (Y.F.)
2 Department of Electrical Engineering, Zhejiang University, Hangzhou 310058, China;
xuzheng007@zju.edu.cn
* Correspondence: yujingqiu@zju.edu.cn; Tel.: +86-178-5731-8731




Received: 30 December 2018; Accepted: 28 January 2019; Published: 30 January 2019


Abstract: The modeling of the Ultra-High Voltage (UHV) tower plays an important role in lightning
protection analysis of transmission lines because the model used will directly affect the reliability
of the results. Moreover, the higher the voltage level is, the more prominent the impact becomes.
This paper first analyzes the inapplicability of the Hara multi-segment multi-surge impedance
model for the ±1100 kV UHV towers, and then builds a non-uniform transmission line model
of the tower. Secondly, the multi-segment multi-surge impedance model is used to study the
influence of the tower’s spatial structure changes on its electromagnetic transient characteristics. It is
concluded that the more accurately the nominal height of the tower is modeled, the more accurately
its electromagnetic transient response is reflected. Finally, the lightning electromagnetic transient
responses of the tower with the non-uniform transmission line model and with the multi-segment
multi-surge impedance model are compared and analyzed, which shows that the non-uniform
transmission line model is more in line with the actual situation under the lightning strikes.

Keywords: ±1100 kV Ultra-High Voltage (UHV) tower; multi-surge impedance model; non-uniform
transmission line model; lightning protection analysis; tower electromagnetic transient response

1. Introduction
Years of experience in line operations at home and abroad have shown that line trip-out accidents
caused by lightning strikes account for 40% to 70% of the total line trip-out accidents [1]. Trip-out
accidents are particularly prominent in areas with complex topography, high soil resistivity or high
lightning density. It can be seen that lightning strikes are still the main threat to the safe and reliable
operation of the transmission lines [2,3]. Due to the complexity of the terrain over long distance
transmission, large span UHV towers are often required. The UHV tower is subjected to a higher level
of overvoltage when struck by lightning than the ordinary height tower. And the insulation level also
requires more. Therefore, the accuracy of the modeling directly affects the reliability of the lightning
protection analysis results of the transmission lines. At present, the multi-surge impedance model is
commonly used in engineering. It can meet the demand of the calculation accuracy of the ordinary
height tower, but cannot meet the requirements for the UHV tower.
When the influence of the earth on the impedance of the vertical conductor is considered, the surge
impedance at different position of the vertical conductor is unequal. The multi-surge impedance model
is based on this theory [4]. The tower model commonly used in engineering is typically the multi-surge
impedance model proposed by Hara et al. [5]. In this multi-surge impedance model, the main body
of the tower is divided into several parts with the cross arms as nodes, and each part is equalized
by a surge impedance. With the increase of the transmission voltage level, the height of the tower

Energies 2019, 12, 445; doi:10.3390/en12030445 www.mdpi.com/journal/energies

Energies 2019, 12, 445 2 of 15

also increases in order to meet the insulation requirements. The nominal height becomes the main
part of the tower, so if only one surge impedance is used to model the nominal height, the influence
of the changes in the spatial structure on the tower’s electromagnetic transient responses cannot be
accurately reflected.
In modelling the UHV tower, based on the Hara multi-surge impedance model, reference [6]
divides the nominal height into 11 segments to study the influence of the tower structure’s changes on
its lightning transient characteristics. In [7], the formula for calculating the multi-wave impedance of
the tower is developed by the theory of biconical antenna and electromagnetic field simulation. In [8],
the non-uniform transmission line model of the tower is derived based on the theory of the biconical
antenna, in which the mutual coupling between different conductors of the tower is considered. In [9],
the non-uniform transmission line model of the composite tower is established and the distributed
parameters of the metal part are also derived based on the theory of biconical antenna. So far there is
little literature about the non-uniform transmission line model of the tower and its applicability in the
UHV towers.
One objective of this paper is to simplify the non-uniform transmission line model for applications
in real engineering. The other objective is to study the applicability of the non-uniform transmission
line model in lightning transient analysis for ± 1100 kV UHV tower. The contributions of this paper
are summarized as follows:

(1) An approach to simplify the non-uniform transmission line model is proposed in this paper.
The vertical parts of the tower are equalized to a single conductor system and its non-uniform
distributed parameters are derived based on the theory of the biconical antenna; and the
horizontal cross-arms are modeled as surge impedances since they are parallel to the ground.
(2) The influence of the tower’s spatial structure changes on its electromagnetic transient response
is studied. It is concluded that the more accurately the nominal height of the tower is modeled,
the more accurately its electromagnetic transient response is reflected.
(3) The applicability of the non-uniform transmission line model in lightning transient analysis
is studied. Through comparing the lightning electromagnetic transient responses with the
non-uniform transmission line model and with the multi-segment multi-surge impedance model,
the applicability of the non-uniform transmission line model is verified.

This paper is organized as follows: in Section 2, the multi-surge impendence model of the tower is
introduced. In Section 3, based on the theory of biconical antenna, the non-uniform transmission line
model of the tower is developed. In Section 4, the influence of the tower’s spatial structure changes
on its electromagnetic transient response is studied by segmenting the nominal height into several
parts firstly; and then the electromagnetic transient responses with the multi-segment multi-surge
impendence model and with the non-uniform transmission line model are compared and analyzed for
the validations of the proposed model. Conclusions are drawn in Section 5.

2. The Multi-Surge Impedance Model of the Tower

The multi-surge impedance model considers both the wave process in the tower and the influence
of the structure and size of the tower on the value of the surge impedance, which is more in line
with the actual situation than the traditional tower model. Common multi-surge impedance models
include multi-layer transmission tower model, multi-layer transmission tower simplified model,
Hara lossless tower model and multi-surge impedance model [5,10–13]. Among them, the surge
response characteristics calculated by the Hara multi-surge impedance model are the closest to the
measured values, shown in Figure 1.
Energies 2019, 12, x FOR PEER 3 of 15

Energies 2019, 12, 445 3 of 15

ZA1 ZA2

Zt1 Zl1
ZA3 ZA4

Zt2 Zl2

Figure 1. The diagram of multi-surge impedance model.

Figure 1. The diagram of multi-surge impedance model.
The multi-surge impedance model consists of a tower body, bracings and cross arms. The Z of
each part is expressed by the formula of its size and geometry. It considers the effect of the cross arms
The multi-surge impedance model consists of a tower body, bracings and cross arms. The Z of
and bracings on the tower surge impedance Z, and is suitable for generalization to the transmission
each parttower
is expressed by the formula of its size and geometry. It considers the effect of the cross arms
with complex structure and high height. This model is mainly used for lightning protection
and bracings on theoftower
simulation 500 kVsurge impedance
transmission Z, and
lines, and is suitable
the surge for generalization
impedance of the tower bodytoneeds
the transmission
to be
tower with complex
corrected structure
to include and high
the influence height.
of the This
initial low modelcoefficient
coupling is mainly used for lightning protection
[14].
The surge impedance of the main legs is given as follows [5]:
simulation of 500 kV transmission lines, and the surge impedance of the tower body needs to be
√ coupling
corrected to include the influence of the initial low ! coefficient [14].
2 2hk
Z = 60 ln
The surge impedance of the maintklegs is given − 2 ( k = 1, 2) (1)
r as follows [5]: ek

 height 
where, hk (k = 1, 2) and rek (k = 1,Z
2) are
tk
the
= 60  ln
2 2hand
k
rek
the (
− 2equivalent )
 k = 1, 2 radius of the kth legs, respectively. (1)
The surge impedance of the bracings  is given by: 
Zlk =and
where, hk (k = 1, 2) and rek (k = 1, 2) are the height 9Ztk (the
k =equivalent
1, 2) (2)
radius of the kth legs, respectively.
The surge impedance of the bracings is given by:
And the surge impedance of the cross arms is given by the following conventional equations for
horizontal conductors: Zlk = 9 Ztk (k2h= 1, 2 )
k
(2)
Z Ak = 60 ln (k = 1, 2, 3, 4) (3)
r Ak
And the surge impedance of the cross arms is given by the following conventional equations for
where, hk is the height between the cross arm and the ground, and rAK is 1/4 of the width of the
horizontalcrossbar
conductors:
at the joint of the tower.

3. Non-Uniform Transmission Line Model of2h 

the Tower
ZAk = 60ln ( k = 1,2,3, 4)
k
(3)
The tower model has a great influence on  the rcalculation
Ak  of lightning overvoltage, so the accuracy
of the results depends on a reasonable tower model. However, in view of the fact that the nominal
where, hk height
is the height
accountsbetween
for almostthethecross
entirearm
partand thetower,
of the ground, and rwith
modeling AK is 1/4
onlyof thesurge
one width of the crossbar
impedance
cannot reflect
at the joint of the tower. the influence of the tower’s spatial structure changes on its transient characteristics
responses. Therefore, considering this deficiency of the multi-surge impedance model, this paper
adopts the non-uniform transmission line model.
3. Non-uniform Transmission Line Model of the Tower
The tower is typically composed of several truss modules, each of which typically includes four
The legs
towerthat model
are connected
has abygreat
other shorter horizontal
influence on the andcalculation
inclined parts.ofThe propagation
lightning of lightning so the
overvoltage,
accuracy of the results depends on a reasonable tower model. However, in view of thewaves.
current on different parts of the tower is a four-dimensional problem involving spherical fact that the
In [15], it is analyzed that less than 1% of the energy of the lightning current propagates along the
nominal inclined
height accounts for almost the entire part of the tower, modeling with only one surge
parts during propagation, and hardly propagates along the horizontal parts. Therefore,
impedance the cannot reflect the
shorter horizontal and influence
inclined parts ofconnecting
the tower’s spatial
the pillar structure
structures may bechanges
disregardedonduring
its transient
characteristics responses.
modelling. Only theTherefore, considering
pillar structure thismodules
of the truss deficiencyneeds ofmodelling,
the multi-surge impedance
that is, the lightning model,
this paper adopts the non-uniform transmission line model.
The tower is typically composed of several truss modules, each of which typically includes four
legs that are connected by other shorter horizontal and inclined parts. The propagation of lightning
current on different parts of the tower is a four-dimensional problem involving spherical waves. In
e which
surgethe impedance
surge impedance of the vertical
of the vertical part
part ofofthe the tower
tower can becan be derived,
derived, and then theand then the
distributed
rsparameters
are derived. The main
are derived. The maintruss
trussmodules
modules ofof thethe
towertower is usually
is usually a trapezoidal
a trapezoidal structure, andstru
its pillar structure
structure is composedis composed of offourfour non-parallel
non-parallel conductors.
conductors. However, when calculating
However, when the calc
parameters of the2019,
Energies infinitesimal
12, 445 cylinder element, we can treat it as four parallel conductor 4 of 15 structures.
rsTheoffour-conductor
the infinitesimal cylinder element, we can treat it as four parallel conductor
system of the infinitesimal cylinder element of each main truss module can be
conductor system
equivalent tocurrent of the
a single-conductor
propagates infinitesimal
structure
only along the legs by the
of the cylinder
clustering
truss element
formula
modules. Therefore, it[8]: ofconsidered
can be each mainthat the truss mod
main mode of the propagation in the tower is the transverse electromagnetic wave (TEM), based on
nt to a single-conductor structure by the clustering formula [8]:
which the surge impedance of the vertical part of the tower can be derived, and then the distributed
4 4 × r × r3
req =modules (4)
parameters are derived. The main truss of thecirc
end tower is usually a trapezoidal structure,
and its pillar structure is composed of four non-parallel conductors. However, when calculating the
r = 4×r ×r
4 we can treat it 3as four parallel conductor structures.
parameters of the infinitesimal cylinder element,
where, req is the equivalent radius of the four- eq conductor system,
end circofrend is the radius of a single conductor
The four-conductor system of the infinitesimal cylinder element each main truss module can be
itself, and rcircequivalent
is the radius of the circumcircle
to a single-conductor ofthe
structure by the four-conductor
clustering formula [8]: system, shown in Figure 2:

is the equivalent radius of the four- conductor 3 system, rend is the radius (4) of a single
q
4
req = 4 × rend × rcirc
r
d rcirc is the radius
where, req is of the circumcircle
the equivalent radius of the four-dof the four-conductor
conductor system, rend is the radius ofsystem,
end
shown in Fig
a single conductor
itself, and rcirc is the radius of the circumcircle of the four-conductor system, shown in Figure 2:

rcirc rend
d

Figure 2. Schematic diagram of the equivalent radius.

The formula for the radius of the circumcircle of the four-conductor system is:
rcirc
Figure 2. Schematic diagram of the equivalent radius.
d d
rcirc = = (5)
Figure 2. Schematic diagram 2sin(πof/ the
The formula for the radius of the circumcircle 4)offour-conductor
the system is:
2 equivalent radius.
d d
rcirc = = √ (5)
2 sin(π/4) 2
If the truss modules are modeled as equivalent conductors
of circular cross section, the tower
formula for the radius of the circumcircle of the four-conductor system is:
can be modeled Ifastheantruss
equivalent
modules arecylindrical network,
modeled as equivalent shownofincircular
conductors Figure 3. section, the tower can
cross
be modeled as an equivalent cylindrical network, shown in Figure 3.

d d L1 L2

rcirc = =
2sin(π / 4)
L3
2 L4 L5

L6
e truss modules are modeled as equivalent conductors of circular cross section
odeled as an equivalent cylindrical network, shown in Figure 3.

Figure 3. Tower on the DC pole and the equivalent model of the tower.
L1 of Lthe
Figure 3. Tower on the DC pole and the equivalent model 2 tower.
Therefore, its electromagnetic transient response can be described by the telegraph equation as
L3
Therefore, its [16]:
follows electromagnetic transient (h,t)
response can be described by the telegraph equation as
− ∂u∂h = R(h)i (h, t) + L(h) ∂i(∂hh,t) L4 L5
follows [16]: (6)
− ∂i(∂hh,t) = C (h) ∂u∂h
(h,t)

where, L(h), R(h), and C(h) ∂are

u( hthe, t )per-unit length inductance,∂resistance,
i( h , t ) and capacitance of the
− varies with
conductor, respectively. They
∂h
( )
= Rthehheight
i( h ,of
t )the ( )
+ Linfinitesimal
h
∂h
L6 element.
cylinder
(6)
∂i( h , t ) ∂u( h , t )
− = C (h)
∂h ∂h

where, L(h), R(h), and C(h) are the per-unit length inductance, resistance, and capacitance of the
Energies 2019, 12, 445 5 of 15

The inductance L(h) consists of the following parts:

L ( h ) = L G ( h ) + L E ( h ) + LC ( h ) (7)

where, LG (h) reflects the magnetic flux in the air, LE (h) reflects that the earth is not an ideal conductive
plane, and LC (h) is the inner inductance of the conductor. The resistance R(h) consists of the
following parts:
R ( h ) = R E ( h ) + RC ( h ) (8)

where, RE (h) reflects the limited conductivity of the earth; RC (h) is the internal resistance of the tower,
accounting for the active power loss when the electromagnetic field penetrates inside the conductor.
As for the capacitance C(h), it is only composed of the geometric capacitance CG (h). It reflects the
electric field effect around the conductor.

3.1. Modeling of the Horizontal Cross Arms

For the horizontal cross arm portion L1 , L2 , L4 and L5 , each is equivalent to a single conductor
and is parallel to the ground, so the parameters of the conductor are equal everywhere, and each
horizontal cross arm can be modeled by one surge impendence. The surge impedance of the cross arm
ZA can be calculated by the following formula [17]:

2hcross
Z A = 60 ln (9)
r
where, hcross is the height between the cross arm and the ground, and r is 1/4 of the width of the
crossbar at the joint of the tower.

3.2. Modeling of the Vertical Main Body

For the vertical parts L3 and L6 of the tower, which are perpendicular to the ground, the parameters
at different heights are different; so the vertical body of the tower is modeled as a non-uniform
transmission line. Since the cross-section radius of the vertical part is much smaller than its length,
the lightning wave can be regarded as a spherical wave propagating on the tower. The propagation
process of the spherical wave is similar to the transverse electromagnetic wave propagation process of
the double-cone antenna.
Figure 4 shows an infinitesimal element dh with a radius req and a height h above the ground on
the cylindrical conductor of length H. And the propagation of the spherical wave in the infinitesimal
element can be regarded as the double-cone antenna shown on the right side of Figure 4. In the figure,
p is the penetration depth of the wave propagating to the earth, and can be calculated as follows [16]:

1
p= p (10)
jωµ E σE

where, µE and σE are the magnetic permeability and electrical conductivity of the soil, respectively,
and ω is the angular frequency of the lightning current.
 req eq

θ1
dh
ies 2019, 12, x FOR PEER 6

Energies 2019, 12, 445 6 of 15

h r
rdθ
dh req req
π 2
θ1
dh p σE ≠∞
H Plane of V=0
h h r
rdθ
dh π 2

p σE ≠∞
σEH≠∞ Plane of V=0
h
(a) (b)

σE ≠∞
e 4. Conical antenna model of vertical conductor infinitesimal element. (a) The infinite
nt of the vertical conductor; (b) (a)
The Conical antenna model.
(b)
Figure 4. Conical antenna model of vertical conductor infinitesimal element. (a) The infinitesimal
main
Figuretruss structure
element
4. Conical ofmodel
the
of the vertical
antenna tower
conductor;
of (b) Theis
vertical usually
Conical trapezoidal,
antenna model.
conductor infinitesimal so the(a)equivalent
element. rad
The infinitesimal
finitesimal
element of theelement
vertical is obtained
The mainconductor;
truss structure (b) The
of the byConical
tower the proportional
is usuallyantenna
trapezoidal,model. method
so the equivalent [9],
radius shown
of the vertical in Figure
infinitesimal element is obtained by the proportional method [9], shown in Figure 5:
The main truss structure of the tower is usually trapezoidal, so the equivalent radius o
cal infinitesimal element is obtained by the proportional method [9], shown in Figure 5:

r0 req r1
h

r0 req r1
h
H

Figure 5. Equivalent radius of the vertical infinitesimal element of the tower.

H
Figure 5.ItsEquivalent radius
equivalent radius is: of the vertical infinitesimal element of the tower.
r0 − r1
req = r0 − ×h (11)
H
Figure 5. Equivalent radius of the vertical infinitesimal element of the tower.
uivalent radius
where, r0 is:
is the equivalent radius of the tower base, r1 is the equivalent radius of the tower top, H is
the height of the tower, and h is the height at which the vertical infinitesimal element is located.
Its equivalent radius is:field and magnetic field strength of a transverse electromagnetic wave propagating in
The electric
r −r
a biconical antenna can be expressed in a spherical coordinate system [18]:
req = r0 −−jβrr0 − r 1 × h
H ϕ = A0 e 0/H
r = p r − (r sin ×h

 1 θ)
eq
Eθ = 0 µ/εH ϕ

H
Er = E ϕ = Hθ = Hr = 0
(12)
is the equivalent radius of the tower base, r is the equivalent radius of the towe


 √ 1
β = ω µε

re, r0 is the equivalent radius of the tower base, r1 is the equivalent radius of the tower top
of the tower, and h is
A0 is the the height at which the vertical
µ and ε infinitesimal element is loc
height of thewhere,
tower, and vector
h is magnetic
the position,
height at β is the phase
which the constant,
vertical are magnetic permeability
infinitesimal
and dielectric constant of the medium above the ground plane respectively. Hr , Hθ , H ϕ and Er ,
element is located
electric field
The electric Efieldand magnetic
andto magnetic field
field strength
andstrength
of
electric fieldof
a transverse
a transverse electromagnetic
in the three electromagnetic wavewave pr
propag
θ , E ϕ refer the magnetic field strength directions of the spherical
ical antenna
biconical antennacancan be beexpressed
expressedininaaspherical coordinate
spherical coordinate system
system [18]:[18]:

 Hϕ==AA0ee−−j β r // ((rrsin
H sinθθ) )
 ϕ 0
Energies 2019, 12, 445 7 of 15

coordinate system respectively. Here, the medium refers to the air, so µ takes the value 4π × 10−7
H/m, while ε takes the value 8.85 × 10−12 F/m.
Therefore, for the infinitesimal element, the voltage U can be obtained by the following
formula [19]:
π
Z2
U= Eθ rdθ (13)
θ1

where, θ 1 is the half of the cone angle, π/2 is the angle between the axis of the cone and the ground,
shown in Figure 4.
From the Ampere loop theorem, the current flowing through the infinitesimal element is
shown below:
Z2π
I= H ϕ req dϕ (14)
0

Substituting for simplification gives:

r   
µ θ
U= A0 e− jβr ln cot 1 (15)
ε 2

I = 2π A0 e− jβr (16)

The characteristic impedance of the cone antenna can be obtained from Equation (15) and
Equation (16): r   
U 1 µ θ
Z0 = = ln cot 1 (17)
I 2π ε 2
According to the geometric relationship, we have:
q
θ
  (h + p)2 + req
2 + (h + p)
cot 1 = (18)
2 req

Substituting Equation (18) into Equation (17), we can get:

√ q
h2 +req 2 +h (h+ p)2 +req
2 +( h + p )
q q
1 µ 1 µ
√
Z0 = 2π ln req + 2π ln
(19)
ε ε h2 +req 2 +h
= ZG + Z E
q
1
µ h2 + req 2 + h
r
ZG = ln (20)
2πε req
q
(h + p)2 + req2 + (h + p)
r
1 µ
ZE = ln q (21)
2π ε h2 + req 2 + h

In turn, the geometric inductance LG of the vertical conductor and the geometric capacitance CG
are shown below: √
LG = µεZG
(22)
CG = µε( LG )−1
Then the generalized inductance is shown below:
√
L EC = µεZE (23)
 Energies 2019, 12, 445 8 of 15

The ground inductance and ground resistance are shown below:

L E = Re{ L EC }
(24)
R E = −ωIm{ L EC }

Since the conductor has a skin effect when flowing through a high-frequency current,
the conductor itself has internal inductance and resistance, and the calculation formulas are
shown below: √
jωµC ρ
ZC = 2πreq
LC =
Im{ ZC } (25)
ω
RC = Re{ ZC }
where, ρ is the resistivity of the conductor; µc is the magnetic permeability of the conductor.
Therefore, the components LG (h), LE (h), LC (h) of L(h) can be obtained from Equation (22), Equations
(24) and (25). The components RE (h), RC (h) of R(h) can be obtained from Equations (24) and (25).
C(h) can be obtained from Equation (25). According to Equations (7) and (8), the non-uniform
distributed parameters L(h), R(h), C(h) of the vertical part L3 and L6 can be obtained. As for the
horizontal cross arms L1 , L2 , L4 and L5 , each horizontal cross arm can be modeled by one surge
impendence, calculated by Equation (9). Thus, the whole non-uniform transmission line model can
be established.

4. Simulation Analysis

4.1. Simulation Model and Parameters

Figure 6 shows the actual shape of the ±1100 kV tower used for the calculation and its model
Energies 2019, 12, x FOR PEER
diagram for this paper. The parameters are shown in Table 1. Among them, the surge impedance of 9 of 15
the main body part Zt , the bracings Zl , and the cross arms ZA can be calculated by Equations (1)–(3),
respectively. R is the tower grounding resistance, which is set to 10 Ω. The calculated surge impedance
of the model is shown in Table 2.

3 9.1 ZA1 ZA2

6.5
A Zt1 Zl1
ZA3 ZA4

20.4
Zt2 Zl2

78

Unit：m
10.73
Figure 6. Parameters and the multi-surge impedance model of the tower.
Figure 6. Parameters and the multi-surge impedance model of the tower.
Table 1. Parameters of the tower.

Table
Tower Parameters 1. Parameters of the tower.
Parameters of the Main Body
Tower material Iron Number of columns 4
TowerTower
Parameters
resistivity 9.09 × 10−7 Ω/m Column Parameters
equivalent radiusof the Main
0.1 m Body
Tower magnetic permeability 4π × 10−4 H/m Maximum distance between columns 10.73 m
Tower materialGround resistivity Iron30 Ω·m Number
Minimum distance of columns
between columns 3.0 m 4
Tower resistivity 9.09 × 10 Ω/m
−7 Column equivalent radius 0.1 m
Tower magnetic permeability 4π × 10 H/m
−4 Maximum distance between columns 10.73 m
Ground resistivity 30 Ω·m Minimum distance between columns 3.0 m

Table 2. Parameters for the multi-surge impedance model.

 Tower magnetic permeability 4π × 10−4 H/m Maximum distance between columns 10.73 m
Ground resistivity 30 Ω·m Minimum distance between columns 3.0 m

Energies 2019, 12, 445

Table 2. Parameters for the multi-surge impedance model. 9 of 15

Parameter Surge Parameter Surge

NameTable 2. Parameters
Impedance/Ω Name Impedance/Ω
for the multi-surge impedance model.
ZA1 288 ZA2 288
Parameter Name Surge Impedance/Ω Parameter Name Surge Impedance/Ω
ZA3 270 ZA4 270
ZA1 288 ZA2 288
ZA3
Z t1
270
180 Z
ZA4
l1 1620
270
Zt1 Zt2 180141 ZZ
l1
l2 1269
1620
Zt2 141 Zl2 1269

The lightning current is simulated by the double-exponential model. The impedance of the
The lightning
lightning channel is 300 Ωcurrent
and its is amplitude
simulated byisthe
260double-exponential
kA. Besides, takemodel.
the topThe
of impedance of the
the tower as a lightning
lightning channel is 300 Ω and its amplitude is 260 kA. Besides, take the top of the tower as a
strike point for the analysis. The insulator is simulated with a 1 pF capacitor and the Leader
lightning strike point for the analysis. The insulator is simulated with a 1 pF capacitor and the Leader
Development Method
Development is taken
Method as its
is taken flashover
as its flashover criterion. The
criterion. The transmission
transmission line connecting
line connecting the two the two
sides of the
sidestower adopts
of the tower the PSCAD
adopts the PSCAD Frequency
Frequency Dependent (phase)
Dependent (phase) model.
model. The distant
The distant towers and
towers and
transmission lines adopts
transmission the Bergeron
lines adopts the Bergeron model
model[20]. Itssurge
[20]. Its surgeimpedance
impedance 300 Ω,300
takestakes andΩ, andspeed
wave wave speed
takes 3 × 108 m/s, shown in Figure 7.
takes 3 × 10 m/s, shown in Figure 7.
8

Lightning current
model

ZA1 ZA2
Distant Frequency Frequency Distant
tower and Dependent Dependent tower and
transmission (phase) Zt1 (phase) transmission
Zl1
line model model ZA3 model line model
ZA4

Insulator model Insulator model

Zt2 Zl2

R
Tower model

Figure 7. Simulation model for this paper.

4.2. Simulation Results of the Multi-Segment Multi-Surge Impedance Model

For the UHV tower, the nominal height is larger, occupying more than 50% of the height of the
tower. The structure and size of the tower body from the top to the bottom vary greatly. If only one
surge impedance is used to model the nominal height, the influence of the spatial structure changes
of the tower on its electromagnetic transient responses cannot be accurately reflected. Therefore,
this paper divides the nominal height into 4, 6, and 8 segments, and analyzes the influence of the
refined division on the electromagnetic transient characteristics of the tower. The improved tower
model is shown in Figure 8. The surge impedance of each segment is calculated by Equations (1)–(3).
The parameters calculated for each model are shown in Tables 3–5 respectively.
of the tower on its electromagnetic transient responses cannot be accurately reflected. Therefore, this
paper divides the nominal height into 4, 6, and 8 segments, and analyzes the influence of the refined
division on the electromagnetic transient characteristics of the tower. The improved tower model is
shown in Figure
Energies 8. 445
2019, 12, The surge impedance of each segment is calculated by Equations10 (1)–(3).
of 15 The
parameters calculated for each model are shown in Tables 3–5 respectively.
ZA1 ZA2 ZA1 ZA2 ZA1 ZA2

Zt1 Zl1 Zt1 Zl1 Zt1 Zl1

ZA3 ZA4 ZA3 ZA4 ZA3 ZA4
Zt2 Zl2
Zt2 Zl2 Zt2 Zl2
Zt3 Zl3
Zt3 Zl3 Zt4 Zl4
Zt3 Zl3
Zt4 Zl4 Zt5 Zl5
Zt6 Zl6
Zt4 Zl4 Zt5 Zl5
Zt7 Zl7
Zt5 Zl5 Zt6 Zl6
Zt8 Zl8
Zt7 Zl7 Zt9 Zl9
R R R

Figure 8. Tower model when the nominal height is divided into 4, 6, and 8 segments.
Figure 8. Tower model when the nominal height is divided into 4, 6, and 8 segments.
Table 3. Parameters for four segments of the nominal height.

Table
Parameter Name3. Parameters for four segments
Surge Impedance/Ω of the
Parameter nominal
Name height.
Surge Impedance/Ω
ZA1 288 A2 Z 288
Parameter
ZA3
Surge
270
Parameter
ZA4
Surge270
Zt1Name Impedance/Ω
180 Name
Zl1 Impedance/Ω
1620
Zt2 ZA1 160
288 ZZA2l2 2881440
Zt3 130 Zl3 1170
Zt4 ZA3 270
96 ZZA4l4 270 864
Zt5 46 Z 414
Zt1 180 Zl1l5 1620
Zt2 160 Zl2 1440
Table 4. Parameters for six segments of the nominal height.
Zt3 130 Zl3 1170
Parameter Name
Zt4 Surge Impedance/Ω
96 Parameter
Zl4 Name Surge Impedance/Ω
864
ZA1 Zt5 28846 ZA2l5 414 288
ZA3 270 ZA4 270
Zt1 180 Zl1 1620
Zt2Table 4. Parameters
163 for six segments of
Zl2the nominal height.
1467
Zt3 143 Zl3 1287
Parameter
Z t4 Surge
121 Parameter
Zl4 Surge
1089
Zt5Name 97
Impedance/Ω Zl5
Name 873
Impedance/Ω
Zt6 67 Zl6 603
Zt7 ZA1 20288 ZZA2
l7
288 180
ZA3 270 ZA4 270
180
Zt1Table 5. Parameters Z l1 1620
for eight segments of the nominal height.
Zt2 163 Zl2 1467
Parameter Name Surge Impedance/Ω Parameter Name Surge Impedance/Ω
Zt3 143 Zl3 1287
ZA1 Zt4 288
121 Z
Zl4
A2 1089288
ZA3 270 ZA4 270
Zt1 Zt5 18097 ZZl5l1 8731620
Zt2 Zt6 16567 ZZl6l2 6031485
Zt3 Zt7 14920 Z
Zl7l3 1801341
Zt4 134 Zl4 1206
Zt5 117 Zl5 1053
Zt6
Table 5. Parameters98for eight segmentsZof
l6 the nominal 882
height.
Zt7 77 Zl7 693
Zt8 48 Zl8 432
Zt9 3 Zl9 27
 Energies 2019, 12, 445 11 of 15

This paper uses PSCAD/EMTDC to calculate the parameters of the multi-surge impedance
model when the nominal height is divided into 1, 4, 6, and 8 segments, and analyzes the transient
characteristics of the tower when lightning strikes. The intersection point of the lightning protection
line and the tower is taken as the lightning strike point. That is point A in Figure 6. The obtained
voltage waveform at the top of the tower is shown in Figure 9, and the dynamic process near the peak
is shown in Figure 10. From the fluctuation process of the voltage in Figure 9 and the peak value of the
voltage Figure 10, we can conclude:

(1) As the number of segments in the nominal height increases, the amplitude of the voltage gradually
decreases. Compared with the one segment model, the overvoltage amplitude of the 8 segments
model decreased by about 15.6%. This is because as the number of the divided segments increases,
the geometric spatial structure of the nominal height is modeled by more surge impedances;
and the impedance values of the respective parts are sequentially reduced, so that the peak value
of the voltage is correspondingly reduced.
(2) When the nominal height is segmented differently, the overall fluctuation trend of the voltage is
almost the same.
(3) As the number of segments of the nominal height increases, the peak of the voltage appears
earlier and earlier. Compared with the one segment model, the time the peak appears is about
0.15 µs ahead of time when the nominal height is divided into 8 segments. This is because the
more the number of segments of the nominal height is, the smaller the height and the larger the
equivalent radius is. Therefore, the surge impedance of the portion closer to the ground is smaller.
That is, when the lightning wave has not transmitted to the bottom of the tower, some of the
reflected waves have returned to the top of the tower. This reflected wave slows down the rise of
the peak voltage and decreases the peak value.
(4) When the number of segments of the nominal height increases, the gradient of the voltage peak
reduction is getting smaller and smaller. For example, when 6 segments or 8 segments are used to
represent the nominal height of the tower, the peak value and fluctuation of the voltage waveform
Energies 2019, 12, at
x FOR PEER
the top of the tower are not much different. It indicates that as the number of segments for 12 of 15
modelling the nominal height reaches a certain level, the peak value will change little and reach its
saturation value. Then, further subdivision will have little effect on the electromagnetic transient
response of the tower.

1 segment
9000
4 segments
7500
6 segments
8 segments
The voltage of A/kV

6000

4500

3000

1500

-1500

-3000
0 5 10 15 20
t/us
Figure 9. Simulation results of the voltage for multi-segment multi-surge impedance model.
Figure 9. Simulation results of the voltage for multi-segment multi-surge impedance model.

10000
1 segment
4 segments
V
 -3000
0 5 10 15 20
t/us

Figure 9. Simulation results of the voltage for multi-segment multi-surge impedance model.
Energies 2019, 12, 445 12 of 15

10000
1 segment
4 segments

The peak value of the voltage/kV

8000 6 segments
8 segments

6000

4000

2000

0
0.0 0.5 1.0 1.5
t/us
Figure 10. The peak value of the voltage for multi-segment multi-surge impedance model.
Figure 10. The peak value of the voltage for multi-segment multi-surge impedance model.
In summary, the more parts of the nominal height is divided into, the more accurately the spatial
structure of the
In summary, the tower
more is modeled,
parts ofandthethe closer theheight
nominal wave process on the tower
is divided into, is to
thethemore
actual accurately
situation. the spatia
However, with the height and structure of the tower varying, the number of the segments for the
structurenominal
of theheight
tower is modeled,
is different and
to reach the the closer
saturation the wave
value. Therefore, process
when practicalon the tower
applications is to the actua
require
situation.high
However, with
accuracy, the the height
multi-segment and structure
multi-surge impedanceof model
the tower varying,
needs to the number
test the number of segmentsof the segments
for the saturation value, which lacks flexibility and reliability.
for the nominal height is different to reach the saturation value. Therefore, when practica
applications require Results
4.3. Simulation high ofaccuracy, the multi-segment
the Non-Uniform multi-surge impedance model needs to test
Transmission Line Model
the number of In segments forthe
order to verify the saturation
accuracy value, which
and effectiveness of the lacks flexibility
non-uniform and reliability.
transmission line model,
this paper compares the electromagnetic transient response of the tower with the non-uniform
4.3. Simulation
Energies Results
transmission
2019, ofmodel
12, xline
FOR the Non-Uniform
PEER Transmission
and with the multi-segment Line Model
multi-surge impedance model. The voltage13 of 15
waveforms at the top of the tower are shown in Figure 11, and the dynamic process near the peak is
In order
shownto verify12.the accuracy and effectiveness of the non-uniform transmission line
in Figure model
this paper compares the electromagnetic
10000
transient response of the tower with the non-uniform
Proposed modelimpedance model. The voltage
transmission line model and with the multi-segment multi-surge
1 segment
waveforms at the top of the tower
8000 are shown in Figure 11, 4and the dynamic process near the peak is
segments
6 segments
shown in Figure 12.
The voltage of A/kV

8 segments 6000

4000

2000

-2000

0 5 10 15 20
t/us
Figure 11. Simulation results of the voltage with the non-uniform transmission line model and with
Figure 11. Simulation
the multi-segment results of
multi-surge the voltage with the non-uniform transmission line model and with
model.
the multi-segment multi-surge model.

10000
Proposed model
1 segment
 0 5 10 15 20
t/us

Figure 11. Simulation results of the voltage with the non-uniform transmission line model and with
the multi-segment
Energies 2019, 12, 445 multi-surge model. 13 of 15

10000
Proposed model
1 segment
4 segments

The peak value of the voltage/kV

6 segments
8000
8 segments

6000

4000

2000

0
0.0 0.5 1.0 1.5
t/us
Figure 12. The peak value of the voltage with the non-uniform transmission line model and with the
Figure 12. The peak
multi-segment value of
multi-surge the voltage with the non-uniform transmission line model and with the
model.
multi-segment multi-surge model.
It can be seen from Figures 11 and 12 that the voltage waveforms with the non-uniform
transmission
It can be line
seenmodel
from has almost11theand
Figures same
12fluctuation trend as waveforms
that the voltage the eight-segment
with multi-surge
the non-uniform
impedance model. However, the differences lie in:
transmission line model has almost the same fluctuation trend as the eight-segment multi-surge
impedance model.
(1) The peak However,
of the the differences
voltage with lie in:transmission line model is lower.
the non-uniform
(2) The voltage waveform with the non-uniform transmission line model fluctuates more, but the
(1) The peak of the voltage with the non-uniform transmission line model is lower.
basic fluctuation trend is consistent.
(2) The voltage waveform with the non-uniform transmission line model fluctuates more, but
In the
theory,
basic as the number
fluctuation ofissegments
trend in the vertical section of the tower increases,
consistent.
the electromagnetic transient response of the tower is modeled more accurately. The fact that
theInvertical
theory, as the
section number ofand
is segmented, segments in the
each section vertical by
is modeled section of the
one surge tower increases,
impedance can be the
electromagnetic
understood thattransient response
each section of theistower
of the tower modeled is modeled moredistributed
with uniform accurately. The fact that
parameters. the
Since thevertical
section is segmented, and each section is modeled by one surge impedance can be understood that
parameters of the non-uniform transmission line model vary with the height, compared with the
multi-segment
each section of themulti-surge impedance
tower is modeled model,
with it is similar
uniform that the
distributed number of
parameters. segments
Since of the
the parameters of
non-uniform transmission line model is close to infinity. Therefore, the wave process
the non-uniform transmission line model vary with the height, compared with the multi-segment is more complex
than the multi-segment
multi-surge impedancemulti-surge
model, itimpedance
is similarmodel,
that and
the the peak value
number is reduced correspondingly,
of segments of the non-uniform
which is also closer to the reality.
transmission line model is close to infinity. Therefore, the wave process is more complex than the
In summary, the non-uniform transmission line model is more accurate than the multi-surge
impedance model. The model reflects that the parameters’ changes with the decrease of the height
and the increase of the equivalent radius of the UHV tower, so it is closer to the transient fluctuation
process of the actual tower. In the lightning protection analysis, the accuracy with the non-uniform
transmission line model is higher, and the simulation results are more reliable.

5. Conclusions
In this paper, the multi-surge impedance tower model is introduced, and the non-uniform
transmission line tower model is established. Then the electromagnetic transient response of the tower
with the multi-segment multi-surge impedance model and with the non-uniform transmission line
model are presented and compared. From the results we can draw the following conclusions:

(1) As the number of segments of the nominal height increases, the peak value of the voltage at
the top of the tower is gradually reduced, and the peak time is gradually decreased. When the
number of segments reaches a certain level, the transient response waveform of the tower will
change little.
Energies 2019, 12, 445 14 of 15

(2) The transient response of the tower with the non-uniform transmission line model is very close to
the saturated response of the multi-segment multi-surge impedance model. It indicates that the
results with the non-uniform transmission line model in lightning protection analysis are more
reliable. So the non-uniform transmission line model should be selected.
(3) It can be seen that the non-uniform transmission line model built in this paper is consistent
with the electromagnetic transient response of the multi-segment multi-surge impedance model.
It shows that the proposed model can reflect the changes of the spatial structure of the tower
more realistically, and is more suitable for the lightning transient analysis of ±1100kV UHV tower
than the multi-surge impedance model.
(4) The accuracy of the tower model determines the reliability of the lightning protection analysis.
As the structure of the tower becomes more and more complicated, the ordinary multi-surge
impedance model is no longer applicable, and a more accurate tower model must be built.
The future modeling of the tower should focus on the physical characteristics. The model of the
tower should be able to better reflect the electric and magnetic field generated by the lightning
current and the tower’s own structure.

Author Contributions: X.G. proposed the kernel of the method; Y.F. derived the expressions and provided the
data; J.Y. and Z.X. performed the simulations and wrote the paper.
Funding: This research was funded by State Grid Corporation of China (0711-16OTL09413001-007).
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Wang, C.J.; Zhu, L.Y.; Ji, S.C.; Zhang, Q. Present and development of lightning protection for HV transmission
lines and substations. Insul. Surge Arresters 2010, 3, 35–46.
2. IEEE Working Group on Estimating Lightning Performance of Transmission Lines. A Simplified Method
for Estimating Lightning Performance of Transmission Line. IEEE Trans. Power Appar. Syst. 1985, PAS-104,
919–932.
3. Li, P.G. A new viewpoint about lightning trip-out of UHV transmission lines. Power Syst. Technol. 2000, 24,
63–65.
4. Xu, Z.H. Lighting Transient Study of Transmission Tower and Analysis of Invaded Wave to Substation; North China
Electric Power University (Hebei): Beijing, China, 2006.
5. Hara, T.; Yamamoto, O. Modeling of a transmission tower for lightning-surge analysis. IEE Proc. Gener.
Transm. Distrib. 1996, 143, 283–289. [CrossRef]
6. Wang, D.J.; Zhou, H.; Chen, J.M.; Lao, J.; Zhu, T.; Bao, J.; Zhang, L.; Wang, J. Design of Multi-Surge Impedance
Model for Ultra-High Transmission Tower and Analysis on Its Transient Characteristic Caused by Lightning
Stroke. Power Syst. Technol. 2007, 31, 11–16.
7. Du, L.; Mi, X.; Xiao, N.Z.; Yu, S.; Yang, Y.; Yang, Q. Equivalent Model of Cross Arm and Inclined Holder of
Transmission Tower. High Volt. Eng. 2012, 38, 3025–3032.
8. Gutierrez R, J.A.; Moreno, P.; Naredo, J.L.; Bermudez, J.L.; Paolone, M.; Nucci, C.A.; Rachidi, F. Nonuniform
transmission tower model for lightning transient studies. IEEE Trans. Power Deliv. 2004, 19, 490–496.
[CrossRef]
9. Huangfu, Y.; Wang, S.; Wang, G.; Xu, W.; Zhang, H. Modeling and Insulation Performance Analysis of
Composite Transmission Line Tower under Lightning Overvoltage. IEEE Trans. Magn. 2015, 51, 1–4.
[CrossRef]
10. Mo, F.J.; Chen, Y.P.; Ruan, J.J. Study on transmission tower models and their lightning performance
calculation. Power Syst. Technol. 2004, 28, 80–84.
11. Liang, Y.M.; Ge, D. Tower model in calculation of lightning protection on multi-circuit transmission line
with same tower. High Volt. Eng. 2006, 31, 76–77.
12. Zhang, Y.; Gao, Y.D.; Du, B.; Shi, W. New tower model in calculation of lightning protection on transmission
line. J. Xi’an Jiaotong Univ. 2004, 38, 365–368.
Energies 2019, 12, 445 15 of 15

13. Yang, Q.; Zhao, J.; Sima, W.X.; Feng, J.; Yuan, T. Lightning back-flashover performance of the Yun-Guang
UHV DC transmission lines. High Vol. Eng. 2008, 34, 1330–1335.
14. Zhang, Y.J.; Sima, W.X.; Zhang, Z.J. Summary of the study of tower models for lightning protection analysis.
High Vol. Eng. 2006, 32, 93–97.
15. Ishii, M.; Baba, Y. Numerical electromagnetic field analysis of tower surge response. IEEE Trans. Power Deliv.
1997, 12, 483–488. [CrossRef]
16. Naredo, J.L.; Moreno, P.; Soudack, A.C.; Marti, J.R. Frequency Independent Representation of Transmission
Lines for Transient Analysis through the Method of Characteristics. In Proceedings of the Joint International
Power Conference Athens Power Tech, Athens, Greece, 5–8 September 1993.
17. Ishii, M.; Kawamura, T.; Kouno, T.; Ohsaki, E.; Shiokawa, K.; Murotani, K.; Higuchi, T. Multistory
transmission tower model for lightning surge analysis. IEEE Trans. Power Deliv. 1991, 6, 1327–1335.
[CrossRef]
18. Jordan, E.C.; Andrews, C.L. Electromagnetic Waves and Radiating Systems. Am. J. Phys. 1968, 19, 477–478.
[CrossRef]
19. Du, L.; Mi, X.; Yang, Y.; Yang, Q. The equivalent model of transmission tower under lightning striking.
High Volt. Eng. 2011, 37, 28–33.
20. Imece, A.F.; Durbak, D.W.; Elahi, H.; Kolluri, S.; Lux, A.; Mader, D.; McDermott, T.E.; Morched, A.;
Mousa, A.M.; Natarajan, R.; et al. Modeling guidelines for fast front transients. IEEE Trans. Power Deliv. 2002,
11, 493–506.

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).